[PATCH net] octeontx2: use arch_extension to enable LSE
flat view
WARM2d
IN NET-NEXT
From: Emanuele Rocca <hidden>
Date: 2026-10-05 16:23:29
Also in:
linux-arm-kernel, linux-crypto, lkml
Subsystem:
marvell crypto driver, marvell octeontx2 physical function driver, networking drivers, the rest · Maintainers:
Srujana Challa, Bharat Bhushan, Sunil Goutham, Geetha sowjanya, Ratheesh Kannoth, Subbaraya Sundeep, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
Queued in net-next as 9142c2f8f84b on 2026-10-08.
Instead of using ".cpu generic+lse", use ".arch_extension lse" to
enable LSE.
The ".cpu generic+lse" directive replaces the assembler's current feature
set with generic+lse, dropping other previously enabled features. In
contrast, ".arch_extension lse" enables LSE while preserving the current
feature set.
For example, the following code is rejected with "selected processor does
not support retaa", because ".cpu generic+lse" drops PAuth even though it
was enabled by armv9-a:
.arch armv9-a
.text
.cpu generic+lse
retaa
We encountered this issue while building the kernel with GCC configured
using --with-arch=armv9-a. Since Armv9-A includes pointer authentication,
GCC can assume that PAuth instructions are available and may use retaa
when generating PAuth-protected returns.
Fixes: caa2da34fd25a ("octeontx2-pf: Initialize and config queues")
Signed-off-by: Emanuele Rocca <redacted>
---
drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h | 2 +-
include/linux/soc/marvell/octeontx2/asm.h | 6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
index eecee612b7b2..c525c557b38e 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h@@ -786,7 +786,7 @@ static inline u64 otx2_atomic64_add(u64 incr, void __iomem *addr) u64 __iomem *ptr = addr; u64 result; - __asm__ volatile(".cpu generic+lse\n" + __asm__ volatile(".arch_extension lse\n" "ldadd %x[i], %x[r], [%[b]]" : [r]"=r"(result), "+m"(*ptr) : [i]"r"(incr), [b]"r"(ptr)
diff --git a/include/linux/soc/marvell/octeontx2/asm.h b/include/linux/soc/marvell/octeontx2/asm.h
index d683251a0b40..a9d29f5d6eeb 100644
--- a/include/linux/soc/marvell/octeontx2/asm.h
+++ b/include/linux/soc/marvell/octeontx2/asm.h@@ -17,7 +17,7 @@ #define otx2_lmt_flush(ioaddr) \ ({ \ u64 result = 0; \ - __asm__ volatile(".cpu generic+lse\n" \ + __asm__ volatile(".arch_extension lse\n" \ "ldeor xzr, %x[rf], [%[rs]]" \ : [rf]"=r" (result) \ : [rs]"r" (ioaddr)); \
@@ -30,7 +30,7 @@ */ #define cn10k_lmt_flush(val, addr) \ ({ \ - __asm__ volatile(".cpu generic+lse\n" \ + __asm__ volatile(".arch_extension lse\n" \ "steorl %x[rf],[%[rs]]" \ : [rf] "+r"(val) \ : [rs] "r"(addr)); \
@@ -40,7 +40,7 @@ static inline u64 otx2_atomic64_fetch_add(u64 incr, u64 *ptr) { u64 result; - asm volatile (".cpu generic+lse\n" + asm volatile (".arch_extension lse\n" "ldadda %x[i], %x[r], [%[b]]" : [r] "=r" (result), "+m" (*ptr) : [i] "r" (incr), [b] "r" (ptr)
--
2.47.3